Temporal alignment of videos based on object viewpoint

The talk addresses the problem of temporally aligning semantically similar videos, for example two videos of cars on different tracks. We present an alignment method that establishes frame-to-frame correspondences such that the two cars are seen from a
similar viewpoint (e.g. facing right), while also being temporally smooth and pleasing to the eye.
Unlike previous works, our method does not assume that the videos show the same scripted sequence of events. We compare against three alternative methods, including the popular DTW algorithm, on a new dataset of realistic videos
collected from the internet. We performed a comprehensive evaluation using a novel protocol that includes both quantitative measures and a user study on visual pleasingness.